Discrepancies between PicoGreen and qPCR in GA libraries

[fedora]

Hi all GA users.
Some ChIP libraries we make are giving a ~4x lower concentration by qPCR compared to PicoGreen. These same libraries when sequenced generate low cluster number (~4x less than other libraries) which isnt increased by using a higher library concentration. The reads we get are high quality and sample DNA-specific. <2% of reads contain the adapter sequence GATCGGAAGA the libraries do not seem to contain the ~120bp adapter multimer band I have seen others comment on. Has anyone observed this sort of thing before or have a potential explanation of the cause? Thanks!

[cjohns]

I'm not sure if this is the answer or not, but for several preps we got huge peaks at 80bp that we believe to be primer dimers. When we put them on the GAII they did not produce any sequences. Maybe this is because they are too short to fold over and get amplified?

[fedora]

Thanks cjohns. We dont see any 80bp or 120b products in these libraries (our DNA is 150-300bp) but maybe there is a causative connection there somewhere. I guess if there were sequences which are complementary to the oligos on the FC but which didnt for some reason contain both amplification primer sequences (or the sequencing primer target) they could mop up the FC oligos and yield nothing. Maybe we need to clone and sequence this stuff...

[Susanne]

This sounds a bit like it could be ssDNA filling up your FC, like you suggested. I've heard of similar things before, although I'm not sure where that amount of ssDNA comes from (preferred amplification of one strand??). That would at least explain the concentration difference you see with qPCR vs. PicoGreen.
However to get rid of ssDNA ?? - I thought the gel extraction should do the job, as ssDNA migrates slower than dsDNA.